1. MOUTON HYCOM developments and tests in the MOUTON project SHOM : Rémy Baraille, Annick Pichon, Gaëlle Casagrande, Cécile Renaudie, (Sébastien Lahaye), Yves Morel Partners : Michel Assenbaum, Sophie Casitas, Stéphanie Corréard, Philippe Craneguy, Nicolas Filatoff, Michel Gavard, Son Hoang

2. MOUTON 2 Objective of the project : Develop and validate a model able to represent the ocean dynamics (=currents, temperature and salinity fields) in both deep and shallow regions at high resolution (1 km) Operational needs : Under water warfare (navig., acoustics in coastal areas) Mine warfare (mine drift + burial ) Special operations (commandos, amphibious ops) Police operations at sea (rescue at sea, drift of objects, pollutions, …). Launched 2001, finish 2008/9 Numerical Modeling : Include tide in OGCMs « Clean » river run-off Deal with the cont. shelf … => based on HYCOM Data Assimilation : Observations (Altimetry, RHF, …) « filters » dev. … Validation : Campaigns at sea CONGAS/MOUTON campaigns 2004->2009

3. MOUTON 3 CHOICE OF THE AREA : Process studies Tide + internal tide Current + drift Upwelling internal waves (solitons) Med waters Coastal/deep ocean exchanges Seasonal stratification (formation of thermocline) Wetting/drying tide O/A intercations (HF) Vortices/Eddies

4. MOUTON 4 Characteristics of the demonstration models V 2.1 : local model – Resolution 1,7 km 33 levels Initial state and boundary conditions forced by MERCATOR MOG2D (tide – Le Provost and Lyard) Atmospheric forcings: ARPEGE (METEO-France) 150 m (only tide) V 2.2 : local model – Resolution 1,7 km V 1 : basin model – Resolution 1/12° assimilation of altimetry (RSMAS+COAPS)

5. MOUTON 5 Numerical developments and results Wetting drying version of HYCOM (modified barotp)  tides New time stepping for the slow part of barotropic fields (slight modif main and momtum)  to deal with strong currents in shallow areas (tides) Time varying stratification characteristics (new subroutine)  manage seasonal thermocline New boundary conditions for BT mode (new subroutine)  clean forcing of the tide and rivers Original data assimilation method  Adaptive filter (adjoint) 4th order advection scheme for momentum (new momtum)  more efficient than 2nd order scheme at high resolution

6. MOUTON 6 Wetting/Drying 20-25 september 1997 Water depth Non-linear terms included in barotropic Mode (no approximations) Drastic modification Of the numerical Schemes (but based On the schemes used In HYCOM for Baroclinic mode)  Has a cost ! (but only for barotropic mode)

7. MOUTON 7 New time stepping for BT mode Baclin = 12 s batrop = 1,5s Baclin = 12 s batrop = 0,5s Baclin = 6 s batrop = 1,5s Baclin = 12 s batrop = 1,5s LSBM Slow evolution of Btropic mode : original code is unstable (CFL) !

8. MOUTON Time evolution of stratification characteristics HYCOM ATL 1/12° AXBT AXBT MOUTON – MAY 2004 HYCOM GASC- MOUTON MERCATOR Position of geop. levels is evolved Monthly according to the averaged stratification in the area from climatology

9. MOUTON HYCOM ATL 1/12° AXBT AXBT MOUTON – JULY 2004 HYCOM GASC-M MERCATOR Time evolution of stratification characteristics

10. MOUTON AXBT MOUTON JULY 2004 (zoom 0-200m) HYCOM GASC-M OBS

11. MOUTON 11 Temperature : comparison between observations (- -) and model (-) (data source : SOMLIT + EPOC) Boundary conditions : rivers

12. MOUTON 12 Salinity : comparison observations (- -) model (-) (data source : SOMLIT + EPOC) Boundary conditions : rivers

13. MOUTON 13 TIDES IN THE MANCHE (CHANEL) model includes all modifications (all are important)

14. MOUTON 14 TIDE M4 Observations (Le Provost/Chabert d’Hières) Model

15. MOUTON 15 TIDE M2 Observations (Le Provost/Chabert d’Hières) Model

16. MOUTON 16 SST – May 2004

17. MOUTON 17 OI : applied to bassin and regional demonstration models Adaptive filter : applied to bassin demonstrator (Indian ocean)  Adjoint of HYCOM Used to deal with high frequency (tides) using HF radars DATA ASSIMILATION SST error (after 5 months) : free run and with assimilation (SSH, SST + in-situ)

18. MOUTON 18 Conclusion : ongoing and future developments Locally adaptive stratification (modification of « hybgen »)  improve evolution of thermocline and prepare coupling with atmospheric models (Cécile Renaudie – SHOM/MF) Simulation over 20 days, wind=20 m/s, air temperature=1°C, radiative flux=0W/m2 with 40 hybrid layers, in May (dp00=1.01m, dp00x=1.41m, dp00f=1.04) compared with an all z-levels solution (4476 layers of 1m each). Alternative : dp00f from the previous time step geometrical series to calculate dp00 :

19. MOUTON 19 Conclusion : ongoing and future developments most important processes (coastal) are reproduced at least qualitatively but need some modification of the code  confidence in model for coastal modeling Data assimilation : Deal with high/low frequency processes (tides) Adapt « adaptive filter » to regional model Pursue validation (test upwellings, gravity currents, internal tides) AGRIF package (modification of « entire » code : look out for syntax in future release)  automatic 1-2 way nesting